Forge Control Plane
One place to direct the infrastructure you run.
Forge gives operators a shared control plane for applications, game servers, databases, network allocations, backups, and the nodes that execute them.
What Forge is
A control plane, not just a panel.
Forge Web gives people a clear interface. Forge API records what should exist, schedules durable work, and coordinates status. PostgreSQL keeps the durable control-plane state. Nodes do the physical execution work through Beacon.
- An operator describes the workload.
- Forge records and coordinates the operation.
- Beacon runs the node action through Docker.
- Health and observations return to Forge.
The node agent.
Beacon runs where workloads run. It has the Docker socket, the workload data directory, console, SFTP, backup adapters, and health reporting. It is not another web panel and it does not schedule independently.
Understand Beacon →Observe before you act.
Forge combines platform health, node observations, container state, operations, logs, and monitoring. Use this evidence before moving work, changing traffic, or making recovery decisions.
See the architecture →Restore from shared evidence.
Recovery is built around a verified backup, eligible destination capacity, and a Beacon that can access shared storage. It is restore-oriented, not automatic ownership failover.
Read recovery guidance →Move deliberately.
Evacuation needs more than a button: drain the source, verify a second healthy Beacon, reserve destination capacity, restore or redeploy, then validate before retiring the source.
Plan an evacuation →The administrator panel
Every section has a job.
The panel groups control-plane decisions by the part of Forge they affect. Some advanced areas are intentionally marked as planning or registry-only where their runtime path is not complete.
See all panel features →Startup guide
Build a safe first workload.
Establish ownership
Create the organization, project, and environment that will own the workload.
Add execution
Enroll a Beacon node, confirm its health, and understand its capacity and storage.
Prepare access
Create a template, allocation, optional mount, and workload database host.
Deploy and observe
Create the app or game server, then watch the operation, logs, and health signals.
Prove recovery
Configure backups and perform a restore test before the workload is important.